CN103955623B - External network measuring and sampling error correction method based on prediction and correction algorithm - Google Patents
External network measuring and sampling error correction method based on prediction and correction algorithm Download PDFInfo
- Publication number
- CN103955623B CN103955623B CN201410206710.XA CN201410206710A CN103955623B CN 103955623 B CN103955623 B CN 103955623B CN 201410206710 A CN201410206710 A CN 201410206710A CN 103955623 B CN103955623 B CN 103955623B
- Authority
- CN
- China
- Prior art keywords
- measuring
- measuring point
- subsystem
- sub
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000005070 sampling Methods 0.000 title claims abstract description 17
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 12
- 238000012937 correction Methods 0.000 title abstract description 8
- 230000000977 initiatory effect Effects 0.000 claims abstract description 8
- 239000000284 extract Substances 0.000 claims description 13
- 238000009826 distribution Methods 0.000 claims description 11
- 238000000605 extraction Methods 0.000 claims description 11
- 238000002715 modification method Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 abstract description 9
- 238000005259 measurement Methods 0.000 description 29
- 101000829367 Homo sapiens Src substrate cortactin Proteins 0.000 description 5
- 102100023719 Src substrate cortactin Human genes 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000009499 grossing Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013213 extrapolation Methods 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 241001123248 Arma Species 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000033772 system development Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an external network measuring and sampling error correction method based on prediction and correction algorithm and belongs to the technical field of operation and control of electric power systems. The method comprises the following steps: after the starting of distributed state estimation calculating, a main subsystem and a slave subsystem are respectively used for acquiring parameters used for computing and measuring section time differences; the slave subsystem is used for sending the acquired parameters used for computing and measuring the section time differences to the main subsystem; the main subsystem is used for computing and measuring the section time differences and sending the measured section time differences to the slave subsystem; the slave subsystem is used for correcting the measured values of all measured points according to the measured section time differences.The external network measuring and sampling error correction method provided by the invention is applied before distributed state estimation calculation, and is used for adjusting measured data of all control centers, participating in calculation, so as to reduce estimation differences of a boundary region. Therefore, all original measuring information, participating in distributed state estimation calculation, of the boundary region of the subsystem are enabled to be consistent basically, the consistency of intranet measuring information of all subsystems and the local measuring information of a calculation initiating system is ensured, and the calculation accuracy is improved.
Description
Technical field
The invention belongs to the operation of power system and control technology field, more particularly, to a kind of being based on are estimated and correcting algorithm
Outer net measure sampling error modification method.
Background technology
At present, one of trend of power system development is exactly Power System Interconnection, and in addition the access of the new forms of energy such as large-scale wind power is right
The impact of network trend and dynamic characteristic is extended to network-wide basis, objectively requires electrical network correlation All Control Center (system) to pass through
Interacted system state estimation sets up the unified online computation model of the whole network.In actual motion, each control centre is only responsible for supervision
The network of oneself affiliated area, is each entered by the real time data that SCADA system obtains local power grid operation according to some cycles
Row state estimation calculates.Because All Control Center state estimation obtains, the data moment is inconsistent each other, dividing between control centre
When cloth state estimation calculates, the time difference is had using the metric data from different control centres, this will bring outer net to estimate
Error, the accuracy of impact state estimation.Because the measurement sampled data from SCADA does not have time scale information, lead to from not
The time difference with the measuring section of control centre is difficult to direct access, thus cannot be directly according to the measuring section time difference external network data
It is modified.
Often there is greater overlap modeling region, using the overlapping modeling of different control centres between actual electric network control centre
Identical measuring point difference this feature of measuring value in region, measures change curve according to each measuring point in overlay region, estimates out two systems real
The border section time difference, other adjacent subsystems were corrected to metric data according to the section time difference, made the institute of participation Distributed Calculation
The measuring section having control centre is under synchronization, improves the accuracy that distributions are estimated.
Document 1 (Zhao Hongga, Xue Yusheng, Gao Xiang, Pan Yongwei, Cen Zonghao, Li Bijun.《The delay inequality of measurement is estimated to state
The impact of meter and its countermeasure [J]》. Automation of Electric Systems, 2004,21:Measurement transmission delay is analyzed to state estimation in 12-16)
The impact of precision, and be uniformly distributed model and propose processing method according to metric data time delay.But the method inapplicable forwarding
The data of data and direct collection the situation deposited.
For measurement transmission delay problem, document 2 (Leou, R.C., Lu, C.N.Adjustment of the External
Network's Measurements and its Effect on the Power Mismatch.Thirty-Forth IAS
Annual Meeting [C] .1999,102 (3):2072-2076) propose two kinds of new forecast models and improve degree of accuracy, respectively
It is winters season autoregression model and ARMA model, but the computational accuracy of the method is difficult to ensure that.
Document 3 (SU C L, LU C N.Interconnected network state estimation using
Randomly delayed measurement s.IEEE Trans on Power Systems, 2001,16 (4):
8702878) suppose that metric data follows specific distribution, and propose Stochastic propagation Kalman filtering algorithm to ask solving time delay
Topic, but the calculating of this algorithm is complicated.
Document 4 (show consideration for, Chen Genjun, Chen Songlin, Li Jiuhu, Tang Guoqing, Yu Erkeng.《Metric data time difference compensation state is estimated
Meter method [J]》. Automation of Electric Systems, 2009,08:Time delay according to different measurements in 44-47), introduces the corresponding time difference
Compensating factor, revises the measurement of time delay using the metric data variable quantity of continuous two data section and the time difference compensation factor,
Measure delay problem to process.The method does not provide the concrete grammar asking for section time delay, but think measurement transmission delay it is known that
And it is translated into compensating factor, introduce state estimation and measure equation.
Content of the invention
It is an object of the invention to, provide and a kind of measure sampling error correction side with the outer net of correcting algorithm based on estimating
Method, using identical measuring point difference this feature of measuring value in different control centres overlapping modeling region, according to each measuring point in overlay region
Measure change curve, estimated out for the two system actual profile time differences, master subsystem again the section estimated out the time difference be sent to adjacent
Other subsystems, other subsystems are corrected to metric data according to the section time difference, make participation distributions estimate meter
The measuring section of all control centres calculated is under synchronization, improves the accuracy that distributions are estimated.
To achieve these goals, technical scheme proposed by the present invention is, a kind of based on the outer net estimated with correcting algorithm
Measure sampling error modification method, it is characterized in that methods described includes：
Step 1：Distributions estimate calculate start after, master subsystem and from subsystem obtain respectively for calculate measure
The parameter of the section time difference；
Described master subsystem is the subsystem initiating to calculate；
Step 2：From subsystem, being used for of obtaining is calculated the parameter of the measuring section time difference and send to master subsystem；
Step 3：The master subsystem computation and measurement section time difference simultaneously sends the described Measure section time difference to from subsystem；
Step 4：From subsystem, the measuring value of each measuring point is revised according to the measuring section time difference.
Described master subsystem obtains and includes for the parameter calculating the measuring section time difference：Master subsystem stores according to itself
The historical metrology section of overlapping region, extracts the measuring value in each measuring point nearest 1 cycle, is designated as the first measuring value Si,1；Ask for each
Measuring point first measures the slope K of change near lineari,1, each measuring point second measure the slope K of change near lineari,2With each measuring point
Projection vertical coordinate S on the first regression straight linei,1Y；
Wherein, i=1,2 ..., n, n are the measuring point number of overlapping region；
The described slope K asking for each measuring point first measurement change near lineari,1Including following sub-step：
Sub-step A1：Master subsystem extracts the nearest p of each measuring point from the historical metrology section of the overlapping region that itself stores
The measuring value in individual cycle；Wherein, p is setting value；
Sub-step A2：Measuring value using each measuring point nearest p cycle predicts the measuring value of each measuring point next cycle；
Sub-step A3：Measuring value by the measuring value in nearest for each measuring point 1 cycle and each measuring point next cycle of prediction
It is connected, obtain each measuring point first and measure change near linear；
Sub-step A4：Calculate the slope K that each measuring point first measures change near lineari,1.
The described slope K asking for each measuring point second measurement change near lineari,2Including following sub-step：
Sub-step B1：Master subsystem extracts each measuring point nearest 2 from the historical metrology section of the overlapping region that itself stores
The measuring value in individual cycle；
Sub-step B2：The measuring value in nearest for each measuring point 2 cycles is connected, obtains each measuring point second measurement change approximately straight
Line；
Sub-step B3：Calculate the slope K that each measuring point second measures change near lineari,2.
Described ask for projection vertical coordinate S on the first regression straight line for each measuring pointi,1YIncluding following sub-step：
Sub-step C1：Nearest p of each measuring point of historical metrology profile extraction of the overlapping region that master subsystem stores from itself
The measuring value in cycle；Wherein, p is setting value；
Sub-step C2：According to the measuring value in each measuring point nearest p cycle, calculate the first regression straight line of each measuring point；
Sub-step C3：The measuring value in nearest for each measuring point 1 cycle is projected on the first regression straight line of this measuring point；
Sub-step C4：Calculate the vertical coordinate S of subpointi,1Y.
Described acquisition from subsystem includes for the parameter calculating the measuring section time difference, is stored according to itself from subsystem
The historical metrology section of overlapping region, extracts the measuring value of the nearest a cycle of each measuring point, is designated as the second measuring value Si,2；Ask for
Projection vertical coordinate S on the second regression straight line for each measuring pointi,2Y；
Wherein, i=1,2 ..., n, n are the measuring point number of overlapping region.
Described ask for projection vertical coordinate S on the second regression straight line for each measuring pointi,2YIncluding following sub-step：
Sub-step D1：Extract the measurement in each measuring point nearest p cycle the historical metrology section of overlapping region from subsystem
Value；Wherein, p is setting value；
Sub-step D2：According to the measuring value in each measuring point nearest p cycle extracted, the second recurrence calculating each measuring point is straight
Line；
Sub-step D3：The measuring value in nearest for each measuring point 1 cycle is projected on the second regression straight line of this measuring point；
Sub-step D4：Calculate the vertical coordinate S of subpointi,2Y.
The described time difference calculating each measuring point specifically includes following sub-step：
Sub-step E1：Make i=1, N1=0, N2=0；
Sub-step E2：When measuring point i meets Ki,2>0 and Si,2Y<Si,1Y, or Ki,2<0 and Si,2Y>Si,1YWhen, then make N1=N1+
1；
When measuring point i meets Ki,2<0 and Si,2Y<Si,1Y, or Ki,2>0 and Si,2Y>Si,1YWhen, then make N2=N2+1；
Sub-step E3：If i >=n, execute sub-step E4；Otherwise, make i=i+1, return sub-step E2；
Sub-step E4：If N1≥N2, then according to formula Δ tj=(Sj,1-Sj,2)/Kj,2Calculate the time difference of measuring point j；If
N1<N2, then according to formula Δ tj=(Sj,1-Sj,2)/Kj,1Calculate the time difference of measuring point j；
Wherein, j=1,2 ..., n, n are the measuring point number of overlapping region.
The computing formula of the described measuring section time difference isWherein, n is the measuring point number of overlapping region.
Described step 4 includes following sub-step：
Sub-step F1：From subsystem, change curve is measured according to each measuring point of subsystem historical metrology cross section correct, specially：
As Δ T12The measuring value in each measuring point of historical metrology profile extraction nearest q cycle, root when >=0, is utilized from subsystem
Predict the measuring value of each measuring point next cycle according to the measuring value extracting, further according to the measuring value in each measuring point nearest q-1 cycle
With the measuring value of each measuring point next cycle of prediction, matching obtains each measuring point and measures change curve f (t)；
As Δ T12<The measuring value in each measuring point of historical metrology profile extraction nearest q cycle when 0, is utilized from subsystem, according to
The measuring value in each measuring point nearest q cycle, matching obtains each measuring point and measures change curve f (t)；
Wherein, q is setting value, and t is to measure the moment；
Sub-step F2：Using formulaRevise the measuring value from subsystem each measuring point i；
For revise after from the measuring section of subsystem measuring point i measuring value；
I=1,2 ..., m, m are the measuring point number of the measuring section from subsystem.
Before the present invention is applied to distributions estimation calculating, to the outer net All Control Center (subsystem) participating in calculating
Metric data carry out integrated regulation, to reduce the estimation difference of borderline region so that all participation distributions estimate
The subsystem borderline region initial measurement information calculating is basically identical it is ensured that each subsystem Intranet measurement information is calculated with initiation
The concordance of the local measurement information of system, improves the accuracy of calculating.
Brief description
Fig. 1 is systematic sampling time difference map；
Fig. 2 is subsystem overlapping region schematic diagram；
Fig. 3 is to measure sampling error modification method flow chart based on the outer net estimated with correcting algorithm；
Fig. 4 is that each measuring point first measures change near linear generation schematic diagram；
Fig. 5 is that each measuring point second measures change near linear generation schematic diagram；
Fig. 6 is emulation platform schematic diagram；
Fig. 7 is example error analyses table；
Fig. 8 is application condition curve chart.
Specific embodiment
Below in conjunction with the accompanying drawings, preferred embodiment is elaborated.It is emphasized that the description below is merely exemplary
, rather than in order to limit the scope of the present invention and its application.
For easy analysis, the present embodiment taking two subsystems (subsystem 1, subsystem 2) interconnection as a example illustrates.Assume
Subsystem 1 is the master subsystem initiating Distributed Calculation, and the respective state estimation of two subsystemses participating in Distributed Calculation calculates
Cycle is consistent, is Δ h.Measuring section is to participate in the measuring value of each measuring point that the subsystem of Distributed Calculation obtains in specific period
Set.All measuring point sampling instants in same section are consistent, and system communication cycle is shorter, measure variation tendency consistent, often
Corresponding 1 collection moment in individual cycle.
Often there is between actual electric network control centre greater overlap modeling region, the measurement calculating each measuring point in overlay region is poor
Value, if measuring mean difference in allowed band, is not required to carry out Predictor Corrector.Due to system communication cycle shorter it is assumed that amount
Survey variation tendency consistent, at short notice, connect the secant being formed using 2 points and to measure change curve in the approximate cycle.With
All measuring point sampling instants in one section are consistent, and system communication cycle is shorter, measure variation tendency consistent.Two System History amounts
Survey and carry out linear regression analyses respectively, made regression straight line is approximately consistent.
Fig. 1 is systematic sampling time difference map.As shown in figure 1, transverse axis is time shafts, after representing that each subsystem obtains electric network data
Periodically carry out the moment of state estimation, the longitudinal axis represents subsystems.In Fig. 1, moment t1 (i) and its measuring section before
Represent the Historic Section being stored in data base, the measuring section of moment t2 (j+1) and t1 (i+1) represents the section not gathered.And
Interval [t1 (i), t2 (j+1)] with I express time, II represents interval [t2 (j+1), t1 (i+1)].M21Represent subsystem 2 in t2
(j-2) measuring section in moment.
With subsystem 1 for the system initiating calculating, i.e. master subsystem.In fig. 2, overlay region D is that master subsystem state is estimated
Meter computation model and the region intersected from subsystem state estimation computation model.A measuring point M is arbitrarily taken in the D of overlay regiond, note survey
Point MdIn M1Measuring value in section is S1, and note measuring value S1 measures regression straight line (the first regression straight line) upslide in master subsystem
The vertical coordinate of shadow is Sd1Y, wherein M1Refer to master subsystem to initiate to calculate the calculating section being.Note measuring point MdIn section M2In measurement
It is worth for S2, remember that measuring value S2 is S measuring the vertical coordinate of the upper projection of regression straight line (second regression straight line) from subsystemd2Y, remember Δ
Sd=S1-S2, wherein M2Be from subsystem master subsystem initiate calculate after, be sent to the measuring section of master subsystem.From son
In system, B area arbitrarily takes a measuring point Mb, remember measuring point MbIn M2Middle measuring value is Sb.
After initiating to calculate, master subsystem 1 obtains section M2Afterwards, bad data, the amount in the measuring section that will obtain are rejected
Measured value carries out preliminary process, by irrational data deletion without.Bad data includes：Do not meet the measurement of Kirchhoff's law
Value, measures the irrational measuring value of Sudden Changing Rate, and generator active power exceedes measuring value of its rated power etc..Reject bad data
Afterwards, master subsystem 1 calculates all measuring points in overlay region in M1And M2In measurement difference, if measure mean difference in allowed band,
Then it is not required to carry out Predictor Corrector, algorithm flow chart is as shown in Figure 3.
In Fig. 3, the method that the present invention provides includes：
Step 1：After distributions are estimated to calculate startup, master subsystem (is not fixing, refers to start the subsystem calculating
System) and the parameter for calculating the measuring section time difference is obtained respectively from subsystem.
After master subsystem initiates to calculate, obtain the parameter for calculating the measuring section time difference.Including：Master subsystem is according to certainly
The historical metrology section of the overlapping region of body storage, extracts the measuring value in each measuring point nearest 1 cycle, is designated as the first measuring value
Si,1.Ask for the slope K that each measuring point first measures change near lineari,1, each measuring point second measure the slope of change near linear
Ki,2With projection vertical coordinate S on the first regression straight line for each measuring pointi,1Y.Wherein, i=1,2 ..., n, n are the survey of overlapping region
Point number.
Because historical metrology section includes the measuring section in each cycle, i.e. the measuring value of each each measuring point of cycle, meter
Calculate after initiating, the measuring value of the nearest a cycle in the Historic Section of sampling is designated as the first measuring value S by master subsystemi,1.
Each measuring point first is measured to the slope K of change near lineari,1, its acquisition process is as follows：
Sub-step A1：The each measuring point of historical metrology profile extraction of the overlapping region that master subsystem stores from itself nearest 4
The measuring value in individual cycle.The present embodiment takes setting value p=4, and the value of p can not be too little, too little can affect prediction effect.
Sub-step A2：Measuring value using 4 nearest cycles of each measuring point predicts the measuring value of each measuring point next cycle.
Used for reference in the present invention winters forecast model, exponential smoothing forecast model, linear extrapolation forecast model,
Several ultra-short term bus load Forecasting Methodology such as gray forecast approach forecast model and load variations value model, is predicted respectively.
Wherein, winters forecast model isFor the predictive value of moment n+j,For
The measurement meansigma methodss of moment n,For the smoothing factor of moment n,Season sex factor for moment n+j.In the present embodiment, j
=4.
Other Forecasting Methodologies are any techniques commonly known, are not the emphasis of present invention protection, therefore in the present invention no longer
Repeat.
Corresponding for each Forecasting Methodology predicting the outcome is designated as X respectively1、X2、X3、X4And X5.In order to improve precision of prediction, comprehensive
Conjunction employs above several method and is finally predicted the outcome.According to the prediction accuracy of various methods, predict the outcome point for it
Join different weights omegai, weight determines that formula is：
Wherein, W=[ω1,ω2,ω3,ω4,ω5], e=[1,1,1,1,1]T.H is 5 × 5 matrixes, is to be divided with 5 kinds of methods
Yu Ce not covariance matrix between gained measuring point predictive value and measuring point actual value.With measuring point MdAs a example, various methods are tried to achieve
Predictive value is assumed to be respectively：X1d、X2d、X3d、X4dAnd X5d, try to achieve weight and be followed successively by ω1d, ω2d, ω3d, ω4dAnd ω5d, then measuring point
MdFinal predictive value：
Xd=ω1dX1d+ω2dX2d+ω3dX4d+ω4dX4d+ω5dX5d(2)
Sub-step A3：Measuring value by the measuring value in nearest for each measuring point 1 cycle and each measuring point next cycle of prediction
It is connected, the straight line obtaining is used for each measuring point in approximate principal and subordinate's two subsystemses overlapping region and measures change curve, this straight line is each measuring point
First measures change near linear.
For any one measuring point, the measuring value in nearest 1 cycle is Si,1, the measuring value of prediction is Xi,d, by this two
Value is connected, and obtains straight line, this straight line is measured change near linear as this measuring point first.As shown in figure 4, measuring point Md's
First measurement change near linear is S1d=K1×t+b1.
Sub-step A4：Calculate the slope K that each measuring point first measures change near lineari,1.Due on straight line 2 points it is known that
It is respectively Si,1And Xi.d, therefore straight slope Ki,1Can try to achieve.
Each measuring point second is measured to the slope K of change near lineari,2, its acquisition process is as follows：
Sub-step B1：The each measuring point of historical metrology profile extraction of the overlapping region that master subsystem stores from itself nearest 2
The measuring value in individual cycle.
Sub-step B2：The measuring value in 2 nearest for each measuring point cycles is connected, the straight line obtaining is used for approximate principal and subordinate two
The each measuring point in system overlapping region measures change curve, and this straight line is that each measuring point second measures change near linear.
For any one measuring point, the measuring value in nearest 1 cycle is Si,1, the measuring value in the 2nd nearest cycle is
Si,0.This two measuring values are connected, obtain straight line, this straight line is measured change near linear as this measuring point second.As
Shown in Fig. 5, measuring point MdSecond measurement change near linear be S2d=K2×t+b2.
Sub-step B3：Calculate the slope K that each measuring point second measures change near lineari,2.Due on straight line 2 points it is known that
It is respectively Si,1And Si,0, therefore straight slope Ki,2Can try to achieve.
For projection vertical coordinate S on the first regression straight line for each measuring pointi,1Y, it is as follows that it asks for process：
Sub-step C1：Nearest 4 of each measuring point of historical metrology profile extraction of the overlapping region that master subsystem stores from itself
The measuring value in cycle.
Sub-step C2：According to the measuring value in each measuring point nearest 4 cycles, calculate the first regression straight line of each measuring point.
Because each measuring point has 4 measuring values (there is a measuring value in each cycle in 4 cycles), therefore can be according to this
4 measuring values, in conjunction with linear regression equation, are calculated the regression straight line of this measuring point, as the first recurrence of this measuring point
Straight line.
Sub-step C3：The measuring value in nearest for each measuring point 1 cycle is projected on the first regression straight line of this measuring point.
Sub-step C4：Calculate the vertical coordinate S of subpointi,1Y.As calculated measuring point MdMeasuring value S1 is in its first regression straight line
On projection vertical coordinate be Sd1Y.
After master subsystem initiates to calculate, obtain the parameter for calculating the measuring section time difference from subsystem.Including from subsystem
The historical metrology section of the overlapping region according to itself storage for the system, extracts the measuring value of the nearest a cycle of each measuring point, is designated as the
Two measuring value Si,2；Ask for projection vertical coordinate S on the second regression straight line for each measuring pointi,2Y.Wherein, i=1,2 ..., n, n are
The measuring point number of overlapping region.
Because historical metrology section includes the measuring section in each cycle, i.e. the measuring value of each each measuring point of cycle.Main
After subsystem initiates to calculate, the measuring value of the nearest a cycle of each measuring point can be extracted from subsystem, be designated as the second amount
Measured value Si,2.
Ask for projection vertical coordinate S on the second regression straight line for each measuring point from subsystemi,2YConcrete acquisition process as follows：
Sub-step D1：Extract the measurement in each measuring point nearest 4 cycles the historical metrology section of overlapping region from subsystem
Value.
Sub-step D2：According to the measuring value in each measuring point nearest 4 cycles extracted, the second recurrence calculating each measuring point is straight
Line.
The measuring value in each 4 cycles of measuring point obtains from subsystem, and because each measuring point has 4 measuring values, because
This can be calculated another regression straight line of this measuring point according to this 4 measuring values, as the second recurrence of this measuring point
Straight line.
Sub-step D3：The measuring value in nearest for each measuring point 1 cycle is projected on the second regression straight line of this measuring point.
Sub-step D4：Calculate the vertical coordinate S of subpointi,2Y.As calculated measuring point MdMeasuring value S2 in its second regression line
The vertical coordinate S of projection on straight lined2Y.
Step 2：From subsystem, being used for of obtaining is calculated the parameter of the measuring section time difference and send to master subsystem.In this step
In rapid, from subsystem by the second measuring value Si,2Vertical coordinate S with subpointi,2YSend to master subsystem.
Step 3：The master subsystem computation and measurement section time difference simultaneously sends the described Measure section time difference to from subsystem.
Master subsystem is according to each measuring point the first measuring value Si,1, each measuring point the second measuring value Si,2, each measuring point first measures and becomes
Change the slope K of near lineari,1, each measuring point second measure the slope K of change near lineari,2, each measuring point is in the first regression straight line
On projection vertical coordinate Si,1YWith projection vertical coordinate S on the second regression straight line for each measuring pointi,2Y, calculate the time difference of each measuring point, tool
Body includes：
Sub-step E1：Make i=1, N1=0, N2=0.
Sub-step E2：When measuring point i meets Ki,2>0 and Si,2Y<Si,1Y, or Ki,2<0 and Si,2Y>Si,1YWhen, then make N1=N1+
1.
When measuring point i meets Ki,2<0 and Si,2Y<Si,1Y, or Ki,2>0 and Si,2Y>Si,1YWhen, then make N2=N2+1.
Sub-step E3：If i >=n, execute sub-step E4；Otherwise, make i=i+1, return sub-step E2；
Sub-step E4：If N1≥N2, then according to formula Δ tj=(Sj,1-Sj,2)/Kj,2Calculate the time difference of measuring point j；If
N1<N2, then according to formula Δ tj=(Sj,1-Sj,2)/Kj,1Calculate the time difference of measuring point j, j=1,2 ..., n, n are overlapping region
Measuring point number.
Then, the time difference according to each measuring point, calculate the measuring section time difference.Its computing formula isWherein, n
For measuring point number in measuring section.
Finally, master subsystem is by Δ T12Send to from subsystem.
Step 4：From subsystem, the measuring value of each measuring point is revised according to the measuring section time difference.Revise the measuring value bag of each measuring point
Include:
Sub-step F1：From subsystem, change curve is measured according to each measuring point of subsystem historical metrology cross section correct, specifically side
Formula is as follows：
As Δ T12The measuring value in each measuring point of historical metrology profile extraction nearest q cycle, root when >=0, is utilized from subsystem
Predict the measuring value of each measuring point next cycle, the Forecasting Methodology that the method for prediction and sub-step A2 provide according to the measuring value extracting
Identical.After obtaining the measuring value of each measuring point next cycle, the measuring value using each measuring point nearest q-1 cycle is each with predict
The measuring value of measuring point next cycle, matching obtains each measuring point and measures change curve f (t).Wherein, q is setting value, and t is measuring point
The collection moment of measuring value.
As shown in figure 1, setting q=4, that is, extract the measuring value in each measuring point nearest 4 cycles.As Δ T12When >=0, boss is
System 1 sampling is ahead of from subsystem 2, now section M24Do not gather, utilize measuring section M from subsystem 220、M21、M22And M23In advance
Survey M24Middle measuring point MbMeasuring value.After predicting new value, using measuring section M21、M22、M23And M24Middle measuring point MbMeasuring value,
Using newton differential technique, matching obtains each measuring point and measures change curve fb(t).Newton differential technique fitting formula is as follows：
F (x)=Pn(x)+Rn(x)
Pn(x)=f (x0)+f[x0,x1](x-x0)+f[x0,x1,x2](x-x0)(x-x1)+... (3)
+f[x0,x1,...,xn](x-x0)...(x-xn-1)
In formula (2),
Rn(x)=f (x)-Pn(x)=f [x0,x1,...,xn]ωn+1(x)
In formula (3), ωn+1(x)=(x-x0)(x-x1)(x-x2)...(x-xn).
As Δ T12<The measuring value in each measuring point of historical metrology profile extraction nearest q cycle when 0, is utilized from subsystem, according to
The measuring value in each measuring point nearest q cycle, matching obtains each measuring point and measures change curve f (t).
In Fig. 1, it is ahead of master subsystem 1 from subsystem 2 sampling, using the existing historical metrology section M of subsystem 221、M22、
M23And M24Middle measuring point MbMeasuring value, according to newton differential technique, matching measures change curve fb(t).
Sub-step F2：Using formulaRevise the measuring value from subsystem each measuring point i.After revising
The measuring value of measuring point i from the measuring section of subsystem, i=1,2 ..., m, m are the measuring point of the measuring section from subsystem
Number.
Embodiment 2
The present invention passes through experiment simulation platform shown in Fig. 6, builds Bulk power system simulation model, simultaneously in EMS on RTDS
(OPEN-3000) set up the physical device model consistent with RTDS phantom in.Illiteracy west electricity by stable operation on RTDS
Pessimistic concurrency control, is split as the electrical network in two regions having overlap, models respectively in EMS system.Passing through wide area network simulator again will
The interconnection of two simulation control centers EMS, simulates different control centres and obtains data moment nonsynchronous state, thus simulation with
In actual electric network, consistent experimental situation is run at intarconnected cotrol center.The present invention set each system state estimation calculating cycle as
30s, the test data actual profile time difference of use is 20s, initiates to calculate with EMS1.
1) estimate part
A) utilize overlay region historical metrology section in EMS1 to predict new measuring section, seek each measuring point amount in two system overlapping regions
Survey change curve：First next cycle measuring section is predicted according to each measuring point historical metrology of EMS1 overlay region, further according to each
The up-to-date historical metrology value of measuring point and predictive value, make each measuring point in the approximate two system overlapping regions of straight line and measure change curve, obtain
Slope, remembers that straight slope required by certain measuring point X is K1.
When predicting next cycle measuring section according to master subsystem historical metrology, the present invention has used for reference winters model, refers to
Number smoothing techniques, several ultra-short term bus load Forecasting Methodology such as linear extrapolation, gray forecast approach, Load Derivation, respectively
Its different weight of distribution that predicts the outcome, comprehensive above several method is predicted.
B) utilize overlay region historical metrology section in EMS1, ask each measuring point in two system overlapping regions to measure change curve.Root
According to each measuring point of master subsystem overlay region up-to-date moment and a upper moment measuring value, make the approximate two system overlapping regions of straight line and respectively survey
Point measures change curve, obtains slope, and straight slope required by note measuring point X is K2.
C) estimate section time difference Δ T12.Using master subsystem historical metrology section, make the first regression straight line y1, obtain calculating
The each measuring point in moment main system overlay region measures the vertical coordinate of projection on the first regression straight line, and it is vertical that note measuring point X measuring value projects
Coordinate is S1Y.Using from subsystem historical metrology section, make the second regression straight line y2, obtain the calculating moment from subsystem overlay region
Each measuring point measures the vertical coordinate of projection on the second regression straight line, and the vertical coordinate of the measuring value projection of note measuring point X is S2Y.For weight
Each of folded area measuring point, can measure variation tendency according to it, and S1Y, S2YSize come to determine master subsystem initiate calculate
Moment is located interval, so that it is determined that each measuring point in overlay region measurement change curve slope is required in should be b).Further each measuring point
Time difference Δ t in two systemsi.Worth Δ T is averaging to the measurement time difference of all measuring points in overlay region12=24.6089 seconds.
2) correction portion
EMS1 is by the section estimated out time difference Δ T12It is sent to EMS2 within=24.6089 seconds.
A) EMS2 historical metrology is utilized to predict new measuring section, each measuring point of matching system 2 measures change curve：Estimate and ask
Obtain Δ T12=24.6089>0, master subsystem 1 sampling is ahead of from subsystem 2, now predicts new measuring section；Using 3 history
Measuring point measuring value in the new section of measuring section and prediction, measures change curve, root with newton each measuring point of differential technique matching
Obtain each measuring point according to the section time difference and repair rear measuring value.
According to prediction correction principle, the measuring value of the overlapping region D of subsystem 2 should be with benchmark system (boss after being corrected
System 1) overlay region D collection real time data consistent, distributions estimate the measurement using each system Intranet after being corrected
Information is carried out.The table that Fig. 7 is given is example error analyses table.This table be the overlapping region D of subsystem 2 is corrected after measure with
Subsystem 1 overlay region D measures and is contrasted, and is 7 measuring points taking at random in measuring section in table.Fig. 8 is by measuring point in form
Error is expressed as broken line, can apparent find out, the D area of the subsystem 2 after estimating-correcting measures and the D calculating moment subsystem 1
Area's measuring value is more closely, algorithm is more accurate.In Fig. 8, dotted line is relative error before adjustment, and solid line misses relatively for before adjustment
Difference.
The above, the only present invention preferably specific embodiment, but protection scope of the present invention is not limited thereto,
Any those familiar with the art the invention discloses technical scope in, the change or replacement that can readily occur in,
All should be included within the scope of the present invention.Therefore, protection scope of the present invention should be with scope of the claims
It is defined.
Claims (9)
1. a kind of measure sampling error modification method with the outer net of correcting algorithm based on estimating, it is characterized in that methods described includes：
Step 1：After distributions are estimated to calculate and are started, master subsystem and obtaining respectively for calculating measuring section from subsystem
The parameter of the time difference；
Described master subsystem is the subsystem initiating to calculate；
Step 2：From subsystem, being used for of obtaining is calculated the parameter of the measuring section time difference and send to master subsystem；
Step 3：Master subsystem calculates the measuring section time difference and sends the described measuring section time difference to from subsystem；
Step 4：From subsystem, the measuring value of each measuring point is revised according to the measuring section time difference；
Described master subsystem obtains and includes for the parameter calculating the measuring section time difference：The overlap that master subsystem stores according to itself
The historical metrology section in region, extracts the measuring value of the nearest a cycle of each measuring point, is designated as the first measuring value Si,1；Ask for each survey
Point first measures the slope K of change near lineari,1, each measuring point second measure the slope K of change near lineari,2Exist with each measuring point
Projection vertical coordinate S on first regression straight linei,1Y；
Wherein, i=1,2 ..., n, n are the measuring point number of overlapping region.
2. method according to claim 1, it is characterized in that described in ask for each measuring point first and measure the oblique of change near linear
Rate Ki,1Including following sub-step：
Sub-step A1：Master subsystem extracts each measuring point nearest p week from the historical metrology section of the overlapping region that itself stores
The measuring value of phase；Wherein, p is setting value；
Sub-step A2：Measuring value using each measuring point nearest p cycle predicts the measuring value of each measuring point next cycle；
Sub-step A3：Measuring value phase by the measuring value of nearest for each measuring point a cycle and each measuring point next cycle of prediction
Even, obtain each measuring point first and measure change near linear；
Sub-step A4：Calculate the slope K that each measuring point first measures change near lineari,1.
3. method according to claim 1, it is characterized in that described in ask for each measuring point second and measure the oblique of change near linear
Rate Ki,2Including following sub-step：
Sub-step B1：Master subsystem extracts each measuring point nearest 2 week from the historical metrology section of the overlapping region that itself stores
The measuring value of phase；
Sub-step B2：The measuring value in nearest for each measuring point 2 cycles is connected, obtains each measuring point second and measure change near linear；
Sub-step B3：Calculate the slope K that each measuring point second measures change near lineari,2.
4. method according to claim 1, it is characterized in that described in ask for projection on the first regression straight line for each measuring point indulge
Coordinate Si,1YIncluding following sub-step：
Sub-step C1：The each measuring point of the historical metrology profile extraction nearest p cycle of the overlapping region that master subsystem stores from itself
Measuring value；Wherein, p is setting value；
Sub-step C2：According to the measuring value in each measuring point nearest p cycle, calculate the first regression straight line of each measuring point；
Sub-step C3：The measuring value of nearest for each measuring point a cycle is projected on the first regression straight line of this measuring point；
Sub-step C4：Calculate the vertical coordinate S of subpointi,1Y.
5. the method according to any one claim in claim 1-4, is characterized in that described acquisition from subsystem is used
Include in the parameter calculating the measuring section time difference, the historical metrology section of the overlapping region being stored according to itself from subsystem, carry
Take the measuring value of the nearest a cycle of each measuring point, be designated as the second measuring value Si,2；Ask for throwing on the second regression straight line for each measuring point
Shadow vertical coordinate Si,2Y；
Wherein, i=1,2 ..., n, n are the measuring point number of overlapping region.
6. method according to claim 5, it is characterized in that described in ask for projection on the second regression straight line for each measuring point indulge
Coordinate Si,2YIncluding following sub-step：
Sub-step D1：Extract the measuring value in each measuring point nearest p cycle the historical metrology section of overlapping region from subsystem；Its
In, p is setting value；
Sub-step D2：According to the measuring value in each measuring point nearest p cycle extracted, calculate the second regression straight line of each measuring point；
Sub-step D3：The measuring value of nearest for each measuring point a cycle is projected on the second regression straight line of this measuring point；
Sub-step D4：Calculate the vertical coordinate S of subpointi,2Y.
7. method according to claim 6, is characterized in that the described calculating measuring section time difference specifically includes following sub-step：
Sub-step E1：Make i=1, N1=0, N2=0；
Sub-step E2：When measuring point i meets Ki,2＞ 0 and Si,2Y<Si,1Y, or Ki,2<0 and Si,2Y＞ Si,1YWhen, then make N1=N1+1；
When measuring point i meets Ki,2<0 and Si,2Y<Si,1Y, or Ki,2＞ 0 and Si,2Y＞ Si,1YWhen, then make N2=N2+1；
Sub-step E3：If i=n, execute sub-step E4；Otherwise, make i=i+1, return sub-step E2；
Sub-step E4：If N1≥N2, then according to formula Δ tj=(Sj,1-Sj,2)/Kj,2Calculate the time difference of measuring point j；If N1<N2,
Then according to formula Δ tj=(Sj,1-Sj,2)/Kj,1Calculate the time difference of measuring point j；
Wherein, j=1,2 ..., n, n are the measuring point number of overlapping region.
8. method according to claim 7, is characterized in that the computing formula of the described measuring section time difference is
Wherein, n is the measuring point number of overlapping region.
9. method according to claim 8, is characterized in that described step 4 includes following sub-step：
Sub-step F1：From subsystem, change curve is measured according to each measuring point of subsystem historical metrology cross section correct, specially：
As Δ T12When >=0, utilize the measuring value in each measuring point of historical metrology profile extraction nearest q cycle from subsystem, according to carrying
The measuring value that takes predicts the measuring value of each measuring point next cycle, the measuring value further according to each measuring point nearest q-1 cycle and in advance
The measuring value of each measuring point next cycle surveyed, matching obtains each measuring point and measures change curve f (t)；
As Δ T12<When 0, utilize the measuring value in each measuring point of historical metrology profile extraction nearest q cycle from subsystem, according to each survey
The measuring value in point nearest q cycle, matching obtains each measuring point and measures change curve f (t)；
Wherein, q is setting value, and t is to measure the moment；
Sub-step F2：Using formulaRevise the measuring value from subsystem each measuring point i；
For revise after from the measuring section of subsystem measuring point i measuring value；
I=1,2 ..., m, m are the measuring point number of the measuring section from subsystem.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410206710.XA CN103955623B (en) | 2014-05-15 | 2014-05-15 | External network measuring and sampling error correction method based on prediction and correction algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410206710.XA CN103955623B (en) | 2014-05-15 | 2014-05-15 | External network measuring and sampling error correction method based on prediction and correction algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103955623A CN103955623A (en) | 2014-07-30 |
| CN103955623B true CN103955623B (en) | 2017-02-08 |
Family
ID=51332898
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410206710.XA Expired - Fee Related CN103955623B (en) | 2014-05-15 | 2014-05-15 | External network measuring and sampling error correction method based on prediction and correction algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103955623B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109752652B (en) * | 2017-11-07 | 2021-02-02 | 上海大郡动力控制技术有限公司 | Phase current sampling method for permanent magnet synchronous motor |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101236216A (en) * | 2008-02-27 | 2008-08-06 | 南京南瑞继保电气有限公司 | Electrical power system metric data time difference compensation state estimation method |
| CN102361353A (en) * | 2011-10-26 | 2012-02-22 | 东北电网有限公司 | Method for aligning section raw data on basis of double time scale delay evaluation |
-
2014
- 2014-05-15 CN CN201410206710.XA patent/CN103955623B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101236216A (en) * | 2008-02-27 | 2008-08-06 | 南京南瑞继保电气有限公司 | Electrical power system metric data time difference compensation state estimation method |
| CN102361353A (en) * | 2011-10-26 | 2012-02-22 | 东北电网有限公司 | Method for aligning section raw data on basis of double time scale delay evaluation |
Non-Patent Citations (6)
| Title |
|---|
| "Adjustment of the External Network’s Measurements and its Effect on the Power Mismatch Analysis";R.C.Leou等;《Thirty-Forth IAS Annual Meeting. Conference Record of the 1999 IEEE》;19991231;第3卷;第2072-2076页 * |
| "Interconnected network state estimation using randomly delayed measurements";Chun-Lien Su等;《IEEE TRANSACTIONS ON POWER SYSTEMS》;20011130;第16卷(第4期);第870-878页 * |
| "Linear minimum variance estimators for systems with bounded random measurement delays and packet dropouts";Shuli Sun;《Signal Processing》;20090731;第89卷(第7期);第1457–1466页 * |
| "基于异步迭代模式的电力系统分布式状态估计方法";张海波等;《电力系统自动化》;20140510;第38卷(第9期);第125-131页 * |
| "量测数据时差补偿状态估计方法";顾全等;《电力系统自动化》;20090425;第33卷(第8期);第44-47页 * |
| "量测量的时延差对状态估计的影响及其对策";赵红嘎等;《电力系统自动化》;20041110;第28卷(第21期);第12-16页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103955623A (en) | 2014-07-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103279804B (en) | The Forecasting Methodology of super short-period wind power | |
| US9178386B2 (en) | State-matrix-independent dynamic process estimation method in real-time for weakly observable measurement nodes without PMU | |
| Celik | Assessing the suitability of wind speed probabilty distribution functions based on wind power density | |
| CN105205570A (en) | Power grid power sale quantity prediction method based on season time sequence analysis | |
| CN102709908B (en) | Loss prediction method for large-scale wind power-accessed power grid | |
| CN105138739B (en) | It is a kind of meter and dead time effect electric system low-limit frequency quick calculation method | |
| CN110417011A (en) | A kind of online dynamic secure estimation method based on mutual information Yu iteration random forest | |
| CN102590685B (en) | Current matching state estimating method of power distribution network | |
| CN103646152A (en) | Electromagnetic transient simulation method of electric system based on matrix exponential | |
| CN103425878B (en) | Power system Quasi dynamic trend and grid operation situation quick calculation method | |
| CN101807798B (en) | Section data integration method for power system safety and stability online analysis | |
| CN106655190A (en) | Method for solving P-OPF (Probabilistic-Optimal Power Flow) of wind power stations | |
| CN108183512A (en) | A kind of reliability estimation method for the electric system for accessing new energy | |
| CN103996147A (en) | Comprehensive evaluation method for power distribution network | |
| CN109102155B (en) | Ultra-short-term node marginal electricity price probability prediction method and system | |
| CN101794998A (en) | Online transient stability analysis method based on concise expression form of electromagnetic power of single generator in multi-machine power system | |
| CN105305439A (en) | Probability dynamic power flow computing method and system in view of input variable correlation | |
| CN106229976A (en) | Transient rotor angle stability situation predictor method based on data-driven | |
| Rossi et al. | Stochastic evaluation of distribution network hosting capacity: Evaluation of the benefits introduced by smart grid technology | |
| CN103473602A (en) | Theoretical line loss data prediction system and prediction method of power grid | |
| CN110299762A (en) | Active distribution network Robust filter method based on PMU near-realtime data | |
| CN104008433A (en) | Method for predicting medium-and-long-term power loads on basis of Bayes dynamic model | |
| CN105023121A (en) | Method for comprehensively evaluating influences imposed on power distribution network by distributed power supply access | |
| Kazemdehdashti et al. | Stochastic energy management in multi-carrier residential energy systems | |
| CN103678798A (en) | Electromagnetic transient simulation method used for power distribution network comprising distributed power sources. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170208 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |